On Saturday 14 January 2023 at 17:40 UTC, a search endpoint flood targeted a retail banking customer in Southeast Asia. The attack peaked at 415.7 million requests per second and lasted 106 minutes. Traffic originated from 2611 autonomous systems in 66 countries, predominantly a headless-browser farm.
| Vector | search endpoint flood |
| Peak | 415.7 million requests per second |
| Duration | 106 min |
| Time to mitigation | 0.403 s |
| Attack traffic reaching origin | 0.033% |
| Legitimate traffic challenged | 0.62% |
Timeline
The attack was preceded by a publicly announced sales event. Request rates on the targeted routes exceeded their hourly baseline by a factor of 701 within 33 seconds. The risk score of participating clients crossed the challenge threshold automatically and proof-of-work difficulty rose with origin load.
What the customer saw
A brief increase in p99 latency of 64 ms during the first minute, then normal service.
Recommendations
- Enable authenticated origin pulls.
- Lower challenge thresholds on authentication endpoints during high-risk events.
- Add a dedicated rate limit for the targeted route.
The billing model is what got our finance team on board, honestly.
How do you avoid challenging uptime monitors and partners?
Nice to read a vendor blog that admits what went wrong.
Clear and practical, thanks.
Solid runbook advice. The DNS-at-2am point hit home.
Thanks! Yes — the risk score and its components are included in every log record.